Project Summary/Abstract In 2016, the Centers for Disease Control and Prevention (CDC) issued guidelines for ?all prescribers to avoid co-prescribing of benzodiazepine (benzo) and opioids?. Other guidelines also advise against benzos for anxiety because safer substitutes exist (serotonin?norepinephrine reuptake inhibitors [SNRIs] and selective serotonin reuptake inhibitors [SSRIs]). Since the 2016 CDC guideline, use of gabapentinoids as an opioid substitute has risen. While recent data showed a decline in opioid prescribing rates after publication of the CDC guideline, a knowledge gap exists in terms of the rates, effectiveness and safety of opioid/benzo substitute prescribing and co-prescribing, especially in populations at high risk of drug toxicity: persons living with disabilities, the elderly and Medicare enrollees in home health care. We now propose to address this gap in knowledge by comprehensively evaluating the toxicity and effectiveness of opioid and benzo substitutes, alone or in combination, compared to opioids and benzos. We anticipate the emergence of previously unrecognized toxicities from drug substitution and co-prescribing. Our Specific Aims are: 1. Assess temporal change in prescribing rates of opioids, opioid substitutes, benzos and benzo substitutes and their co-prescribing among commercially insured and Medicare patients, and the role of provider, patient, federal and state policy on drug substitution and co-prescribing. 2. Examine toxicities (e.g., falls, fractures, emergency room visits) related to substitutes for opioids and benzos, and how toxicities vary with different combinations of drug substitution and co-prescribing, using longitudinal data analytic methods as well as data mining methods for previously unrecognized toxicities. 3. Examine changes in physical function and measures of pain and anxiety in Medicare enrollees prescribed substitutes for both opioids and benzos, and how these outcomes vary with different combinations of drug substitution and co-prescribing among Medicare enrollees in home health care. We will use 20% national Medicare data to identify the elderly and the disabled Medicare populations, and the Clinformatics Data Mart to identify commercially insured populations. Our overall hypotheses is that the substitute drugs, alone or in combination, have considerably lower rates of serious toxicity than do opioids and benzos, with similar effectiveness in many clinical situations. If this is the case, its demonstration by our proposed research should result in acceleration in the shift away from opioids and benzos to safer alternatives.